agentleFS
zanglgzanglg/dotagents/skills/engflow-research/SKILL.md

engflow-research

Skill0 starsChanged 17 days agoMITView on GitHub

Investigate technical questions through source code, first-party specifications and existing evidence. Return grounded conclusions, counterevidence and uncertainty. Use on explicit research request. Other Engflow entrypoints reuse the shared source-investigation method without invoking this Skill. Save a report only when explicitly requested; no experiment execution or production implementation.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use bounded read-only helpers when they materially improve the investigation, following delegation guidance.

    line 42
  2. Distinctive:

    Do not modify source or configuration, run builds, tests, reproductions or benchmarks, instrument systems, or create/execute PoCs.

    line 50
  3. Distinctive:

    Follow project instructions and effective decisions.

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Engflow Research

Answer the user's technical question with a defensible explanation. Choose the investigation strategy and answer structure to fit the question; no predefined research category, workflow or document setup is required.

Engflow is installed as a complete group; this entrypoint is independently usable on an explicit Research request. Read and apply the shared source-investigation method before investigating. Other entrypoints read that method directly in their own tasks; they do not automatically invoke this Skill or acquire its report-write scope. Explicit support requests can return scoped findings to a caller. Honor host invocation guards and current access/disclosure restrictions.

Results and persistence

Return findings to the caller or, when independent, directly in conversation. Save a report only when explicitly requested and within write authority. Reuse the same appropriate record when continuing a question; preserve useful history and distinguish prior evidence from new observations. Follow the supplied path or project convention; otherwise save to docs/research/NNN-short-name.md.

For a new record, use the next number above allocated research records, starting at 001 with at least three digits. Do not fill gaps, renumber existing records or overwrite a collision; recheck the destination before writing. This numbering is independent of features and PoCs. Preview does not create files or reserve numbers. Saving existing findings requires no rerun or new investigation. Reference relevant feature/design sources without creating them or changing current-feature navigation. The report outline is optional; report its saved path and limits.

Use bounded read-only helpers when they materially improve the investigation, following delegation guidance. Otherwise work directly. The agent responsible for this Research task handles explicitly authorized report writes; investigation-only helpers return evidence. Determine that responsibility from the assigned task and write scope, not merely from being a host-level subagent.

Boundaries

Do not modify source or configuration, run builds, tests, reproductions or benchmarks, instrument systems, or create/execute PoCs. Analyze existing results with their provenance and limitations; suggest runtime verification when needed, but leave execution to separately scoped work.

Follow project instructions and effective decisions. Findings are advisory, not design approval or authority to implement changes. Do not commit, push, switch branches or create/manage worktrees.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.